Great point! We A/B tested a bunch of different ones, we haven't managed to outperform the current one. We'll certainly keep testing it. I like your hypothesis!